Kling 4.0 Test Shows Professional Shots, but Delivery Still Needs Human Backup
奇点研究社 used Kling 4.0 to create an exhibition tour video with live-action footage, AI presenters and seven generated segments, supported by roughly 3,000 Chinese characters of prompts. Some shots achieved professional quality, but the test also exposed incorrect subject placement, broken long takes, unstable edits, voice mismatches and garbled audio. The results indicate that Kling 4.0 can produce usable shots while still requiring manual intervention for dependable end-to-end delivery.
The production test provides concrete evidence of both Kling 4.0’s professional potential and its remaining reliability gaps, particularly where consistent editing, sound and complex scene relationships are required.
（本文作者为 奇点研究社，钛媒体经授权发布）
文 | 奇点研究社，作者｜徐婷，编辑｜奇博士
10月30日深夜，快手在港交所发布公告：旗下视频生成业务可灵完成增资交割，198.02 亿元增资款全部到账，约合 29.05 亿美元，可灵投后估值约 179 亿美元，折合人民币超过 1200 亿元，高于快手当日约 968 亿元的港股收盘市值。
“一个可灵，比整个快手还贵”，尤其考虑到，2026 年上半年，可灵收入占快手总收入的比例仅为 2.17%。资本市场买的显然不是可灵当下的收入，而是它在 AI 视频赛道里的未来。
资本故事讲得再大，最终仍要回到创作现场接受检验。最近，我们用可灵4.0制作一条探展视频，真人出镜环节全部交给 AI，采用展会实拍素材与 AI 人物结合的方式，拆分生成 7 组画面，再通过音画同出、视频修改和时间线等功能完成后续制作。测试前，我们准备了约 3000 字提示词，尽可能明确主体关系、镜头运动、构图和声音要求，减少指令歧义。